Correspondence Response to Letter Regarding Article, “Discontinuation of Smokeless Tobacco and Mortality Risk After Myocardial Infarction”
Ideally, the effects of snus quitting after a myocardial infarction should be studied in the setting of a randomized, controlled clinical trial. In the meantime, our cohort study provides the best available evidence for those effects.
First of all, we thank Drs Rodu and Phillips for contributing to further understanding of the effects of snus in the setting of an acute coronary syndrome. We are happy to provide more details about our study.1 The sample includes few snus-using women. We could therefore not study the effects of quitting snus in men and women separately. Nevertheless, the exclusion of subgroups should be based on subject matter knowledge of differential effects, for which we had none regarding snus and sex. When excluding women, reassuringly, quitting snus was associated with a hazard ratio of 0.54 (95% confidence interval, 0.30–0.96) and 0.56 (0.31–1.01) for models A and B, respectively. This is nearly identical to the observations in the primary study sample. Unfortunately, a typo occurred regarding the total number of person-years at risk, despite the fact that the manuscript had been scrutinized by all coauthors and 4 reviewers. The correct number is 43 070 person-years at risk in the total sample (not 40 370, as mistakenly stated), rendering the reported incidence rate of 18.9 (deaths per 1000 person-years at risk) based on 812 deaths. This typo had no influence on the interpretation of the current study. The question about the tobacco nonusers is interesting. First, the mentioned number of dual users (n=934) was at the time of the myocardial infarction, whereas the number of dual users at baseline (2 months after the myocardial infarction) was lower (n=299). To extrapolate and compare rates based on the previous number is incorrect, because all other rates (all age-adjusted) are defined at baseline. Second, the comment reflects ignorance about the sufficient component cause model, proposed in 1976.2 This is a cornerstone of modern causal inference understanding. Different individuals have different component causes that contribute to the sufficient cause of a myocardial infarction, and for death after a myocardial infarction, as well. Snus may be 1 component cause for snus-using individuals. Removing that component cause may prevent deaths. Nonusers of snus have other component causes of death after myocardial infarction than snus users, involving both measured and unmeasured risk factors. The comparison of snus quitters with nonusers will therefore not contribute to the understanding of snus as a potential cause of death after myocardial infarction.
Disclosures None. Gabriel Arefalk, MD Department of Medical Sciences Uppsala University Uppsala, Sweden Kristina Hambraeus, MD Department of Cardiology Falu Hospital Falun, Sweden Lars Lind, MD, PhD Department of Medical Sciences Uppsala University Uppsala, Sweden Karl Michaëlsson, MD, PhD Department of Surgical Sciences Uppsala University Uppsala, Sweden Bertil Lindahl, MD, PhD Johan Sundström, MD, PhD Department of Medical Sciences Uppsala University Uppsala, Sweden
References 1. Arefalk G, Hambraeus K, Lind L, Michaëlsson K, Lindahl B, Sundström J. Discontinuation of smokeless tobacco and mortality risk after myocardial infarction. Circulation. 2014;130:325–332. doi: 10.1161/ CIRCULATIONAHA.113.007252. 2. Rothman KJ. Causes. Am J Epidemiol. 1976;104:587–592.
(Circulation. 2015;131:e423. DOI: 10.1161/CIRCULATIONAHA.114.014029.) © 2015 American Heart Association, Inc. Circulation is available at http://circ.ahajournals.org
DOI: 10.1161/CIRCULATIONAHA.114.014029
e423
Copyright of Circulation is the property of Lippincott Williams & Wilkins and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
Ideally, the effects of snus quitting after a myocardial infarction should be studied in the setting of a randomized, controlled clinical trial. In the meantime, our cohort study provides the best available evidence for those effects.
First of all, we thank Drs Rodu and Phillips for contributing to further understanding of the effects of snus in the setting of an acute coronary syndrome. We are happy to provide more details about our study.1 The sample includes few snus-using women. We could therefore not study the effects of quitting snus in men and women separately. Nevertheless, the exclusion of subgroups should be based on subject matter knowledge of differential effects, for which we had none regarding snus and sex. When excluding women, reassuringly, quitting snus was associated with a hazard ratio of 0.54 (95% confidence interval, 0.30–0.96) and 0.56 (0.31–1.01) for models A and B, respectively. This is nearly identical to the observations in the primary study sample. Unfortunately, a typo occurred regarding the total number of person-years at risk, despite the fact that the manuscript had been scrutinized by all coauthors and 4 reviewers. The correct number is 43 070 person-years at risk in the total sample (not 40 370, as mistakenly stated), rendering the reported incidence rate of 18.9 (deaths per 1000 person-years at risk) based on 812 deaths. This typo had no influence on the interpretation of the current study. The question about the tobacco nonusers is interesting. First, the mentioned number of dual users (n=934) was at the time of the myocardial infarction, whereas the number of dual users at baseline (2 months after the myocardial infarction) was lower (n=299). To extrapolate and compare rates based on the previous number is incorrect, because all other rates (all age-adjusted) are defined at baseline. Second, the comment reflects ignorance about the sufficient component cause model, proposed in 1976.2 This is a cornerstone of modern causal inference understanding. Different individuals have different component causes that contribute to the sufficient cause of a myocardial infarction, and for death after a myocardial infarction, as well. Snus may be 1 component cause for snus-using individuals. Removing that component cause may prevent deaths. Nonusers of snus have other component causes of death after myocardial infarction than snus users, involving both measured and unmeasured risk factors. The comparison of snus quitters with nonusers will therefore not contribute to the understanding of snus as a potential cause of death after myocardial infarction.
Disclosures None. Gabriel Arefalk, MD Department of Medical Sciences Uppsala University Uppsala, Sweden Kristina Hambraeus, MD Department of Cardiology Falu Hospital Falun, Sweden Lars Lind, MD, PhD Department of Medical Sciences Uppsala University Uppsala, Sweden Karl Michaëlsson, MD, PhD Department of Surgical Sciences Uppsala University Uppsala, Sweden Bertil Lindahl, MD, PhD Johan Sundström, MD, PhD Department of Medical Sciences Uppsala University Uppsala, Sweden
References 1. Arefalk G, Hambraeus K, Lind L, Michaëlsson K, Lindahl B, Sundström J. Discontinuation of smokeless tobacco and mortality risk after myocardial infarction. Circulation. 2014;130:325–332. doi: 10.1161/ CIRCULATIONAHA.113.007252. 2. Rothman KJ. Causes. Am J Epidemiol. 1976;104:587–592.
(Circulation. 2015;131:e423. DOI: 10.1161/CIRCULATIONAHA.114.014029.) © 2015 American Heart Association, Inc. Circulation is available at http://circ.ahajournals.org
DOI: 10.1161/CIRCULATIONAHA.114.014029
e423
Copyright of Circulation is the property of Lippincott Williams & Wilkins and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
